Machine tool management system based on big data analysis

Abstract

The invention discloses a machine tool management system based on big data analysis, relates to the technical field of machine tool management and maintenance, and solves the technical problem that the maintenance efficiency and the data acquisition amount cannot be effectively balanced when regular maintenance or predictive maintenance is performed on a machine tool in the prior art. A fault prediction model is obtained through simulation test data training, and fault sequences corresponding to a plurality of target machine tools are predicted according to the fault prediction model; calculating conventional maintenance cost according to the plurality of fault sequences, and setting a maintenance point when the conventional maintenance cost is greater than the cost required for setting the maintenance point; under the condition that business secrets of clients are not violated, fault prediction is completed only through operation data and working condition environments of the machine tool, and it can be predicted that a certain type of faults possibly occur on the machine tool; according to the fault prediction result, communication with clients can be carried out in advance, maintenance points are set or maintenance personnel are dispatched to maintain a plurality of machine tools, normal work of the machine tools is not affected, and the maintenance efficiency can be effectively improved.

Description

Technical Field

[0001] This application belongs to the field of machine tool management and maintenance, specifically a machine tool management system based on big data analysis. Background Art

[0002] Machine tool maintenance is a key measure to ensure long-term stable operation of machine tools, extend service life, improve processing efficiency and ensure processing quality. Manufacturers mainly carry out maintenance through a combination of regular maintenance, remote monitoring and maintenance, preventive maintenance and professional technical services, aiming to reduce machine tool failures and extend service life.

[0003] When manufacturers conduct remote monitoring and predictive maintenance on machine tools, they obtain the operating data of machine tools through various remote monitoring methods, and then use intelligent diagnostic models to deeply mine and analyze the operating data to identify or predict possible faults of machine tools, and formulate relevant maintenance measures based on the faults. However, if accurate fault prediction is required, a lot of detailed data is required, which is difficult for customers to accept; and the regular maintenance method is difficult to adapt to the changing working conditions of machine tools, which will reduce the efficiency and quality of maintenance.

[0004] This application provides a machine tool management system based on big data analysis to solve the above technical problems. Summary of the invention

[0005] The present application aims to solve at least one of the technical problems existing in the prior art; to this end, the present application proposes a machine tool management system based on big data analysis, which is used to solve the technical problem that when performing regular maintenance or predictive maintenance on machine tools in the prior art, the relationship between maintenance efficiency and data collection volume cannot be balanced, which may lead to the leakage of customers' business secrets.

[0006] To achieve the above-mentioned purpose, the first aspect of the present application provides a machine tool management system based on big data analysis, including a cloud analysis module and a data acquisition module connected thereto;

[0007] Data acquisition module: used to pre-process the operating data of several target machine tools according to the set data acquisition rules to obtain desensitized data; associate the desensitized data with the environmental data corresponding to the target machine tools to obtain target data; wherein the pre-processing includes desensitization and encryption, and the environmental data includes the working environment and location;

[0008] Cloud analysis module: used to simulate and train the fault prediction model, and predict the fault sequence of the target machine tool based on the fault prediction model and target data; wherein the fault sequence includes the fault type and the fault time; and,

[0009] It is used to calculate the regular maintenance cost according to the fault sequences of several target machine tools; determine whether to set up a maintenance point according to the regular maintenance cost; if yes, set up the maintenance point; if not, dispatch maintenance personnel to perform temporary maintenance on the target machine tools; wherein, there is at least one maintenance personnel stationed at the maintenance point.

[0010] Preferably, the simulation training of the fault prediction model includes:

[0011] A fault association model is constructed according to the simulated test data, and a plurality of training data groups are extracted from the fault association model; wherein the training data group includes the test data group and the corresponding fault sequence;

[0012] The artificial intelligence model is trained by a plurality of training data groups to obtain a fault prediction model; wherein the artificial intelligence model is constructed based on a BP neural network model or a RBF neural network model.

[0013] Preferably, constructing a fault association model based on simulated test data includes:

[0014] Setting several groups of simulation test data; simulating the operating state of the target machine tool under the several groups of simulation test data, and converting the operating state into a fault sequence; wherein the simulation test data includes operating data and environmental data;

[0015] Several groups of simulated test data and corresponding fault sequences are input into the constructed Bayesian network model to obtain a fault association model.

[0016] Preferably, the conventional maintenance cost is calculated according to the failure sequences of several target machine tools, including:

[0017] An evaluation cycle is set, and several time periods are extracted from the evaluation cycle in combination with the fault sequences of several target machine tools; wherein the evaluation cycle includes a month, a quarter or a year;

[0018] The period maintenance costs of several time periods are calculated based on several fault sequences, and the regular maintenance costs are obtained by adding up the period maintenance costs. Among them, the period maintenance costs include labor costs and equipment costs.

[0019] Preferably, several time periods are extracted from the evaluation cycle in combination with the fault sequences of several target machine tools, including:

[0020] The fault time in the fault sequence is used as the independent variable, and the fault type is classified and processed as the dependent variable to obtain the fault change diagram; wherein the classification processing includes the classification and numerical processing of the fault type;

[0021] Several time periods are extracted from the evaluation period according to the intervals before and after each fault type in the fault change diagram; wherein the intervals before and after are determined according to the dispatching rules of temporary maintenance personnel.

[0022] Preferably, calculating the maintenance costs of several time periods based on several fault sequences includes:

[0023] Calculate the basic maintenance cost corresponding to each fault type in each time period according to the fault sequence; the basic maintenance cost includes labor cost, equipment cost and seasonal cost;

[0024] The basic maintenance costs corresponding to each fault type in the time period are added up to obtain the period maintenance cost of each time period.

[0025] Preferably, several target machine tools are determined, including:

[0026] Several original ranges are delineated based on administrative regions from large to small; administrative regions include provincial, municipal, town / village and village levels in order;

[0027] When there is no maintenance point in the original range, the original range is marked as the target range; and the machine tool in the target range is marked as the target machine tool.

[0028] Preferably, determining whether to set a maintenance point based on conventional maintenance costs includes:

[0029] Calculate the maintenance facility costs corresponding to the operating maintenance points in the administrative areas corresponding to several target machine tools;

[0030] When the cost of conventional maintenance is greater than the cost of maintenance facilities, a maintenance point is set up; otherwise, temporary maintenance is carried out on the target machine tool by dispatching maintenance personnel.

[0031] Preferably, the cost of maintaining the facilities is calculated, including:

[0032] Determine the attributes of maintenance personnel and maintenance equipment according to the fault type in the evaluation period; wherein the attributes of maintenance personnel include the type of maintenance personnel and the number of maintenance personnel, and the attributes of maintenance equipment include the type of maintenance equipment and the number of maintenance equipment;

[0033] The maximum cost required to set up a maintenance point in the administrative area is calculated based on the attributes of the maintenance personnel and the maintenance equipment, and is marked as the maintenance facility cost.

[0034] Preferably, maintenance points are set, including:

[0035] determining a number of candidate locations;

[0036] Based on the fault sequences of several target machine tools and the distances between the corresponding positions of the several target machine tools and several candidate positions, an optimal candidate position is determined; and a maintenance point is set at the optimal candidate position.

[0037] Compared with the prior art, the beneficial effects of this application are:

[0038] The present application obtains a fault prediction model through training of simulated test data, and predicts the fault sequences corresponding to several target machine tools according to the fault prediction model. The data used for prediction is desensitized data and will not infringe on the customer's business secrets. The conventional maintenance cost is calculated according to several fault sequences. When the conventional maintenance cost is greater than the cost required to set up maintenance points, maintenance points are set to reduce the machine tool maintenance cost. Under the condition of not infringing on the customer's business secrets, fault prediction is completed only through the machine tool's operating data and working environment, and it can be predicted how long it will take for the machine tool to have a certain type of failure. According to the fault prediction results, communication with the customer can be made in advance, and maintenance points can be set or maintenance personnel can be dispatched to maintain several machine tools, which not only does not affect the normal operation of the machine tools, but also can effectively improve the maintenance efficiency.

[0039] After determining that a maintenance point needs to be set up, the present application determines several candidate locations, calculates the maintenance efficiency and rental costs of the several candidate locations as maintenance points, determines the candidate location with the highest maintenance efficiency and the lowest rental cost within the constraints as the optimal candidate location, and sets the maintenance point at the optimal candidate location; the present application comprehensively considers the maintenance efficiency and operating costs to determine the optimal location for the maintenance point, which can reduce the operating costs of the maintenance point while ensuring the maintenance efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 This is a schematic diagram of the system structure of a machine tool management system according to Embodiment 1 of the present application;

[0042] Figure 2 This is a schematic diagram of the working principle of the machine tool management system according to the first embodiment of the present application;

[0043] Figure 3 This is a schematic diagram of the calculation principle of conventional maintenance costs in Example 2 of the present application;

[0044] Figure 4 This is a schematic diagram of the principle of dividing several time periods according to the fault change diagram in the second embodiment of the present application;

[0045] Figure 5 This is a schematic diagram of the process of determining maintenance points in Example 3 of the present application. DETAILED DESCRIPTION

[0046] The technical solution of the present application will be described clearly and completely in conjunction with the embodiments below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0047] Machine tool manufacturers need to maintain the machine tools they sell, such as regularly dispatching maintenance personnel to customers to maintain the machine tools, or dispatching maintenance personnel to perform maintenance after customers proactively contact the manufacturer. Of course, some manufacturers use big data technology to collect and analyze data on the machine tools they sell, predict possible machine tool failures, and then dispatch maintenance personnel to handle them in a timely manner. This method has a high accuracy in machine tool fault diagnosis, but it mainly relies on various detailed data of the machine tools, including the customer's business secrets. It is difficult for customers to accept this method, which also makes it impossible to apply this method in this scenario.

[0048] Embodiment 1: This example provides a machine tool management system based on big data analysis. The machine tool management system predicts the fault of the target machine tool through a fault prediction model trained by simulation combined with the non-confidential data of the target machine tool. Maintenance measures can be taken in time according to the predicted fault. Remote judgment can be completed without obtaining the customer's business secrets, which has higher maintenance efficiency and a wider range of applications.

[0049] See also Figure 1-Figure 2 , the first aspect of the present application provides a machine tool management system based on big data analysis, including: a cloud analysis module, and a data acquisition module connected thereto;

[0050] Data acquisition module: used to pre-process the operating data of several target machine tools according to the set data acquisition rules to obtain desensitized data; associate the desensitized data with the environmental data corresponding to the target machine tools to obtain target data; wherein the pre-processing includes desensitization and encryption, and the environmental data includes the working environment and location;

[0051] Cloud analysis module: used to simulate and train the fault prediction model, and predict the fault sequence of the target machine tool based on the fault prediction model and target data; wherein the fault sequence includes the fault type and the fault time; and,

[0052] It is used to calculate the regular maintenance cost according to the fault sequences of several target machine tools; determine whether to set up a maintenance point according to the regular maintenance cost; if yes, set up the maintenance point; if not, dispatch maintenance personnel to perform temporary maintenance on the target machine tools; wherein, there is at least one maintenance personnel stationed at the maintenance point.

[0053] The core of this embodiment is that the relevant operation data of the target machine tool is not collected, and only the target machine tool is predicted to fail and the type of failure based on the operation data and environmental data. The operation data here is filtered by the data collection rules, and the customer's commercial secret data, such as the frequency of use, the length of use, etc., will not be collected. Of course, since the collected operation data itself is relatively conventional, it can also be collected through a specific sensor set on the machine tool, so there is no need to set data collection rules.

[0054] This embodiment does not use the customer's commercial secret data when predicting the fault of the target machine tool. The prediction can be roughly completed only by the above-mentioned operation data. For example, by simulating the operation status of the same model machine tool under different operation data, the corresponding fault can be obtained. However, the operation of the machine tool is greatly affected by the operation environment, and the environmental data is not commercial secret data. Therefore, the environmental data is introduced to improve the prediction accuracy of the machine tool fault.

[0055] After predicting the fault sequences of several target machine tools, we can know that each target machine tool will (probably) have a certain fault in a certain period of time in the future. The maintenance cost of temporarily dispatching maintenance personnel is calculated according to each fault type and the location of the target machine tool, that is, the regular maintenance cost. If the regular maintenance cost is higher than the cost of setting up a maintenance point, the maintenance cost can be reduced by setting up a maintenance point. Otherwise, maintenance personnel can be directly dispatched for maintenance.

[0056] In this embodiment, the target machine tool can be determined according to the administrative region, including: demarcating several original ranges from large to small based on the administrative region; marking the original range as the target range when there is no maintenance point in the original range; marking the machine tool in the target range as the target machine tool. It is worth noting that the target machine tool can also include a machine tool that is not currently installed but is determined to be installed in the future (sold but not installed).

[0057] Administrative regions include provincial, municipal, town / village and village levels. Several original ranges can be defined based on the machine tools in each administrative region, that is, each original range corresponds to an administrative region, and each administrative region has machine tools. If there is no maintenance point in the original range, the original range is marked as the target range.

[0058] It should be noted that the division according to administrative regions is conducive to the management of machine tool maintenance. Of course, it is also possible to break through the restrictions of administrative regions, such as taking a number of relatively concentrated machine tools as target machine tools. If a number of machine tools in administrative region A and the adjacent administrative region B are relatively concentrated, these machine tools will be taken as target machine tools. It is even possible to re-divide machine tools that have already matched maintenance points to reduce maintenance costs.

[0059] Without infringing on the customer's business secrets, it is also necessary to ensure that the fault prediction model can accurately predict the fault of the target machine tool under the available data. This embodiment obtains the fault prediction model by simulation training, including: building a fault association model based on simulated test data, extracting several training data groups from the fault association model; training the artificial intelligence model through several training data groups to obtain the fault prediction model.

[0060] The content type of the simulation test data is the same as that of the target data, but they differ in value and quantity. Several groups of simulation test data are set in advance, each of which includes operation data and environmental data. Then, based on the several groups of simulation test data, simulation tests are performed on the same model of machine tools to obtain the operation status under each simulation status. The fault sequence of the machine tool is extracted and integrated from the operation status, that is, each group of simulation test data corresponds to a set of fault sequences.

[0061] If the artificial intelligence model is directly trained using simulated test data and corresponding fault sequences, a large number of simulated tests will be required, consuming a lot of resources and increasing the overall cost. However, insufficient data will lead to insufficient accuracy of the trained artificial intelligence model. Therefore, after obtaining several sets of simulated test data and corresponding fault sequences, the relationship between data and faults is deeply mined through the Bayesian network to construct a fault association model.

[0062] Bayesian networks can handle uncertainty problems well in this scenario. In many cases, the cause of machine tool failure is uncertain. Bayesian networks can model and analyze these uncertainty problems through probabilistic reasoning. In addition, Bayesian networks can deeply mine data relationships, so that complex relationships behind data can be discovered without a large number of simulation tests, thereby achieving data expansion. The construction and training of Bayesian networks can refer to existing solutions, and the specific process will not be repeated here.

[0063] From the fault association model constructed by the Bayesian network, the values ​​corresponding to each node can be extracted to generate training data, and a large amount of training data can be extracted to generate a training data set. Each node in the fault association model corresponds to a variable, and the variables include operating data, environmental data, and fault type. In order to reduce the amount of data processing during the construction of the fault association model and the training of the artificial intelligence model, the difference between the environmental data and the standard working environment of the machine tool can be used as the node value when constructing the fault association model, which can effectively reduce the value corresponding to the node and reduce the subsequent data processing volume.

[0064] It is worth noting that both operating data and environmental data are related to machine tools, but do not involve the customer's trade secrets. Operating data mainly includes the operating time and number of operations of the machine tool, while environmental data refers to the working environment of the machine tool, such as temperature, humidity, altitude, etc.

[0065] A large amount of training data is extracted from the fault association model to form a training data group, and the artificial intelligence model is trained through the training data group to obtain a fault prediction model. The artificial intelligence model can be a variety of models suitable for the scenario, such as BP neural network model, RBF neural network model, etc. The model training process can refer to the existing training scheme, which will not be repeated here.

[0066] In some other preferred embodiments, if the customer allows access to relevant commercial secret data and the manufacturer takes confidentiality measures, the data types in the training data set can be increased to maximize the accuracy and reliability of the fault prediction model.

[0067] In this embodiment, without infringing on the customer's commercial secrets, fault prediction is completed only through the machine tool's operating data and working environment, and it can be predicted how long it will take for the machine tool to have a certain type of fault. According to the fault prediction results, it is possible to communicate with the customer in advance, set up maintenance points or dispatch maintenance personnel to perform maintenance on several machine tools, which not only does not affect the normal operation of the machine tool, but also effectively improves the maintenance efficiency.

[0068] Embodiment 2: Based on embodiment 1, this embodiment calculates the cost of dispatching maintenance personnel for temporary maintenance according to several fault sequences, that is, the regular maintenance cost. If the regular maintenance cost is too high, setting up maintenance points is considered to reduce the maintenance cost.

[0069] See also Figure 3 , calculating the conventional maintenance cost according to the fault sequences of several target machine tools, including: setting an evaluation period, extracting several time periods from the evaluation period in combination with the fault sequences of several target machine tools; calculating the maintenance cost of several time periods based on several fault sequences, and superimposing the maintenance costs of several time periods to obtain the conventional maintenance cost.

[0070] To calculate the routine maintenance costs of several target machine tools, it is necessary to determine the time range of the calculation, that is, to pre-set the evaluation period. The evaluation period includes months, quarters or years, which can be reasonably set according to the difficulty of setting maintenance points and the manufacturer's planning. For example, if the manufacturer arranges the maintenance points for the next year at the end of the year, the evaluation period is set to years; if the maintenance points are relatively simple to set, the evaluation period can be set to months.

[0071] This embodiment calculates the routine maintenance cost in time periods because the number of target machine tools is limited and it is difficult to ensure that machine tool failures are evenly distributed within the evaluation period. In addition, the number of dispatches and types of dispatches of maintenance personnel need to be considered when calculating the routine maintenance cost.

[0072] Combined with the fault sequences of several target machine tools, several time periods are extracted from the evaluation cycle, including: taking the fault time in the fault sequence as the independent variable, classifying the fault type as the dependent variable, and obtaining a fault change diagram; extracting several time periods from the evaluation cycle according to the intervals before and after each fault type in the fault change diagram.

[0073] The above classification of fault types mainly classifies machine tool faults according to the type of maintenance personnel, that is, the fault types that can be handled by the same type of maintenance personnel are classified into one category. This is because the machine tool fault corresponding to the target machine tool may require a specific type of maintenance personnel to handle it. When calculating the regular maintenance cost later, the dispatch cost of a certain type of maintenance personnel can be used for calculation.

[0074] An example of obtaining a fault change diagram is as follows:

[0075] Several fault sequences are classified to obtain several types of machine tool faults, and each type of machine tool fault is marked, such as all fault types in category 1 are marked as 1, or marked as 1.1, 1.2, etc. in sequence;

[0076] Taking time as the independent variable and the fault type mark as the dependent variable, a fault change diagram of this type of fault can be obtained. The fault change diagram shows the fault distribution of several target machine tools during the evaluation period.

[0077] According to the fault change graph, it can be determined in which time periods the fault prediction of the target machine tool is concentrated, and the time period can be determined by setting the time interval. For example, if the time corresponding to any machine tool failure is used as a reference, if there is a machine tool failure in front or behind it within the set time interval, the time range corresponding to the two machine tool failures is used as a time period. It can be understood that the time interval between machine tool failures in a time period is not greater than the set time interval, and there is no machine tool failure between adjacent time periods.

[0078] like Figure 4 This is a simple fault change diagram. The lines perpendicular to the horizontal axis in the first quadrant represent machine tool faults. Different heights represent different fault types, which can also be understood as different marks for the above fault types. The entire evaluation cycle can be divided into five time periods according to the time interval before and after the fault. There are three different machine tool faults in time period 1, three identical machine tool faults in time period 2, and so on. In this way, several time periods can be extracted from the evaluation cycle, and the maintenance cost of each time period can be calculated (calculated separately according to the fault type and then superimposed).

[0079] The time intervals set above are mainly determined by the dispatch rules of temporary maintenance personnel. If two similar machine tool failures occur one day apart, the dispatched temporary maintenance personnel will complete the task only after they have dealt with both machine tool failures; if two similar machine tool failures occur one week apart, two temporary maintenance personnel will be dispatched to deal with the two machine tool failures respectively. The labor costs paid by the manufacturer in these two cases are different.

[0080] After extracting several time periods from the evaluation cycle, the machine tool maintenance cost corresponding to each time period is calculated to obtain the time period maintenance cost of each time period. The specific calculation steps can be referred to as follows: Calculate the basic maintenance cost corresponding to each fault type in each time period according to the fault sequence; Add up the basic maintenance cost corresponding to each fault type in the time period to obtain the time period maintenance cost of each time period.

[0081] Basic maintenance costs include labor costs, equipment costs and seasonal costs, which are actually the superposition of the three.

[0082] The calculation formula of human cost HC is: HC = W × H × (1 + B); where W is the average salary of maintenance personnel, H is the total working hours of maintenance personnel, and B is the welfare cost ratio of maintenance personnel. It should be noted that the average salary of maintenance personnel is related to the type of maintenance personnel (how many machine tool failures can be handled). For different machine tool failures in the same time period, if high-level maintenance personnel can handle high-level and low-level machine tool failures, the average salary of maintenance personnel is shared by different machine tool failures; if maintenance personnel of each level are only responsible for machine tool failures of the corresponding level, the average salary of maintenance personnel corresponding to each failure type is calculated separately.

[0083] The calculation formula of equipment cost EC is: EC = (P + M + D) × D f ; Among them, P is the purchase cost of maintenance equipment, M is the maintenance cost of maintenance equipment, D is the depreciation cost of maintenance equipment, D f It is the inverse of the useful life of the maintenance equipment (indicating the impact of annual depreciation). It should be noted that the equipment cost here refers to the maintenance equipment used by maintenance personnel when performing machine tool maintenance. If maintenance equipment is not required, this cost is 0.

[0084] The calculation formula for seasonal cost SC is: Among them, C base represents the basic cost, S j is the cost adjustment ratio of the jth seasonal or temporal factor, T j is the duration ratio of the jth seasonal or temporal factor.

[0085] S jIndicates the impact of the jth seasonal or temporal factor on the cost. This parameter is used to quantify the cost changes caused by environmental changes, market demand fluctuations, changes in human resource costs, etc. in a specific season or time period. For example:

[0086] During the winter months, energy consumption may increase in some areas, leading to higher energy costs.

[0087] In the summer, high temperatures may require additional cooling measures, increasing cooling costs.

[0088] During holidays or weekends, labor costs may increase due to overtime pay.

[0089] S j The value of can be positive or negative, depending on whether the factor increases or decreases costs. This ratio can be determined through historical data analysis, market research, or expert experience.

[0090] T j is the duration ratio of the jth seasonal or temporal factor in the total time period. This parameter is used to reflect the relative importance of a specific season or time period in the entire consideration period. For example:

[0091] If a region has 3 months in winter and there are 12 months in a year, then T in winter is j That is 3 / 12=0.25.

[0092] If an industry is particularly busy in the fourth quarter, then the T j will be relatively high, reflecting the greater need for resources and attention during this time.

[0093] T j The value of is usually between 0 and 1, indicating the proportion of the factor in the total time period. This proportion can help manufacturers understand the importance of resource allocation and cost control in different time periods.

[0094] In some other preferred embodiments, the maintenance cost of each fault type may also be calculated with reference to existing public solutions, or directly set based on experience.

[0095] After obtaining the basic maintenance costs of each fault type within the time period, the maintenance costs for the time period can be obtained by superimposing them. The maintenance costs for the time period are the costs required to dispatch maintenance personnel within the corresponding time period. The maintenance costs corresponding to several time periods are superimposed to obtain the conventional maintenance costs, which is the cost paid by the manufacturer to dispatch maintenance personnel to perform maintenance on several target machine tools during the evaluation period. If the conventional maintenance cost is high and higher than the cost of setting up maintenance points, it is inclined to set up maintenance points; on the contrary, if the conventional maintenance cost is lower than the cost of setting up maintenance points, it is inclined to dispatch maintenance personnel to perform temporary maintenance on the faults of several target machine tools.

[0096] The calculation of the cost of setting up the above maintenance points (maintenance facility cost) can refer to the following steps:

[0097] Determine the attributes of maintenance personnel and maintenance equipment according to the fault type in the evaluation period; calculate the maximum cost required to set up a maintenance point in the administrative area based on the attributes of maintenance personnel and maintenance equipment, and mark it as the maintenance facility cost.

[0098] The attributes of maintenance personnel include the type and number of maintenance personnel, which mainly refer to the types and number of maintenance personnel that need to be dispatched to maintain a number of target machine tools during the evaluation period; the attributes of maintenance equipment include the type and number of maintenance equipment, which mainly refer to the maintenance equipment required for the maintenance of a number of target machine tools and the number of times it is used. These labor costs and equipment costs are estimable costs, but the rental cost of the location corresponding to the maintenance point is difficult to estimate. It can be calculated based on the maximum rental cost corresponding to the administrative area, or it can be calculated based on the average rental cost.

[0099] This embodiment divides the evaluation period into several time periods according to several fault sequences, calculates the total cost of machine tool fault maintenance in each time period, and adds up the total costs of each time period to obtain the conventional maintenance cost. Finally, the conventional maintenance cost is compared with the cost of setting up maintenance points to determine whether to set up maintenance points, which can be used by manufacturers for decision-making.

[0100] Embodiment 3: Based on Embodiment 2, a solution for setting maintenance points is provided, which improves maintenance efficiency by setting maintenance points at reasonable locations to ensure low costs.

[0101] See also Figure 5 To set the maintenance points, refer to the following steps:

[0102] Determine multiple candidate locations in administrative areas corresponding to the target machine tools, wherein the candidate locations are determined by comprehensively considering the cost and the distance from each target machine tool;

[0103] Based on the fault sequences of several target machine tools and the distances between the corresponding positions of the several target machine tools and several candidate positions, an optimal candidate position is determined; and a maintenance point is set at the optimal candidate position.

[0104] The above optimal candidate position can be obtained by formula Among them, x is the candidate position, t k represents the number of faults of the kth target machine tool, w k is the fault weight of the kth target machine tool, which can be set according to the urgency and maintenance difficulty of the machine tool fault. f is the position set of the target machine tool, d(x·P fk ) is the distance between the candidate position and the kth target machine tool, C r (x) is the rental cost of candidate location x. The rental cost is determined according to the local market conditions and geographical location. Accurate data can be obtained through surveys or database queries. x (It can be solved by machine learning algorithm) is to select the x corresponding to the minimum PM among all possible x, that is, the optimal candidate location. The goal of the above calculation formula is to minimize the sum of the products of weighted maintenance demand and distance, while considering the rental cost. By solving the formula, the candidate location that achieves the best balance between cost and maintenance efficiency can be obtained.

[0105] It should be noted that the distance d(x·P fk ) can indicate the accessibility and response speed of maintenance services. The shorter the distance, the shorter the time it takes for maintenance personnel and equipment to reach the faulty machine tool, and the higher the service efficiency, so the larger the reciprocal of the distance. Conversely, the longer the distance, the lower the service efficiency, and the smaller the reciprocal of the distance. The ratio of the weighted fault number and distance of the kth target machine tool reflects the urgency of the machine tool fault and the accessibility of the maintenance service. The target machine tool here refers to the machine tool with faults, which can be obtained by eliminating the machine tools without faults from several target machine tools.

[0106] When determining the optimal candidate location through the above formula, constraints should be set, that is, the range of service efficiency and rental cost, which can be set based on the manufacturer's maintenance experience and acceptable cost. The service efficiency and rental cost of the optimal candidate location should meet the requirements of the constraints.

[0107] In some other embodiments, personnel costs and equipment costs may also be introduced into the formula.

[0108] This embodiment calculates the maintenance efficiency and rental cost of several candidate locations as maintenance points, selects the candidate location corresponding to the best balance point between maintenance efficiency and rental cost as the optimal candidate location, and sets the maintenance point at the optimal candidate location. This can reduce the operating cost of the maintenance point while ensuring the maintenance efficiency.

[0109] The above embodiments are only used to illustrate the technical method of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the preferred embodiments, a person of ordinary skill in the art should understand that the technical method of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present application.

Claims

1. A machine tool management system based on big data analysis, including a cloud analysis module and a data acquisition module connected thereto; characterized in that: Data acquisition module: used to pre-process the operating data of several target machine tools according to the set data acquisition rules to obtain desensitized data; associate the desensitized data with the environmental data corresponding to the target machine tools to obtain target data; wherein the pre-processing includes desensitization and encryption, and the environmental data includes the working environment and location; Cloud analysis module: used to simulate and train the fault prediction model, and predict the fault sequence of the target machine tool based on the fault prediction model and target data; wherein the fault sequence includes the fault type and the fault time; and, Used to calculate the regular maintenance cost according to the fault sequence of several target machine tools; determine whether to set up a maintenance point according to the regular maintenance cost; if yes, set up the maintenance point; if not, dispatch maintenance personnel to perform temporary maintenance on the target machine tools; wherein, there is at least one maintenance personnel stationed at the maintenance point.

2. The machine tool management system based on big data analysis according to claim 1, characterized in that: The simulation training fault prediction model includes: Constructing a fault association model according to the simulated test data, and extracting a number of training data groups from the fault association model; wherein the training data group includes the test data group and the corresponding fault sequence; The artificial intelligence model is trained by using several of the training data groups to obtain a fault prediction model; wherein the artificial intelligence model is constructed based on a BP neural network model or a RBF neural network model.

3. The machine tool management system based on big data analysis according to claim 2 is characterized in that: The method of constructing a fault correlation model according to the simulated test data includes: Setting several groups of simulation test data; simulating the operating state of the target machine tool under the several groups of simulation test data, and converting the operating state into a fault sequence; wherein the simulation test data includes operating data and environmental data; Several groups of simulated test data and corresponding fault sequences are input into the constructed Bayesian network model to obtain a fault association model.

4. The machine tool management system based on big data analysis according to claim 1, characterized in that: The conventional maintenance cost is calculated according to the failure sequences of several target machine tools, including: Setting an evaluation cycle, and extracting a number of time periods from the evaluation cycle in combination with a number of fault sequences of the target machine tools; wherein the evaluation cycle includes a month, a quarter or a year; The period maintenance costs of several time periods are calculated based on the several fault sequences, and the regular maintenance costs are obtained by adding the several period maintenance costs; wherein the period maintenance costs include labor costs and equipment costs.

5. The machine tool management system based on big data analysis according to claim 4 is characterized in that: A plurality of time periods are extracted from the evaluation cycle in combination with the fault sequences of the plurality of target machine tools, including: The fault time in the fault sequence is used as an independent variable, and the fault type is classified and processed as a dependent variable to obtain a fault change diagram; wherein the classification processing includes classification and numerical processing of the fault type; A number of time periods are extracted from the evaluation cycle according to the intervals before and after each fault type in the fault change diagram; wherein the intervals before and after are determined according to the dispatching rules of temporary maintenance personnel.

6. The machine tool management system based on big data analysis according to claim 4 or 5, characterized in that: The maintenance costs of several time periods are calculated based on the several fault sequences, including: Calculate the basic maintenance cost corresponding to each fault type in each time period according to the fault sequence; the basic maintenance cost includes labor cost, equipment cost and seasonal cost; The basic maintenance costs corresponding to each fault type in the time period are added up to obtain the period maintenance cost of each time period.

7. The machine tool management system based on big data analysis according to claim 6, characterized in that: Determine a number of target machine tools, including: Several original ranges are delineated based on administrative regions from large to small; administrative regions include provincial, municipal, town / village and village levels in order; When there is no maintenance point in the original range, the original range is marked as a target range; and the machine tools in the target range are marked as target machine tools.

8. The machine tool management system based on big data analysis according to claim 4, characterized in that: Determine whether to set maintenance points based on the conventional maintenance costs, including: Calculate the maintenance facility costs corresponding to the operating maintenance points in the administrative regions corresponding to the target machine tools; When the conventional maintenance cost is greater than the maintenance facility cost, a maintenance point is set up; otherwise, temporary maintenance is performed on the target machine tool by dispatching maintenance personnel.

9. The machine tool management system based on big data analysis according to claim 8, characterized in that: Calculate the cost of the maintenance facilities, including: Determine the maintenance personnel attributes and maintenance equipment attributes according to the fault type in the evaluation cycle; wherein the maintenance personnel attributes include the maintenance personnel type and the number of maintenance personnel, and the maintenance equipment attributes include the maintenance equipment type and the number of maintenance equipment; The maximum cost required to set up a maintenance point in the administrative area is calculated based on the attributes of the maintenance personnel and the maintenance equipment, and is marked as the maintenance facility cost.

10. The machine tool management system based on big data analysis according to claim 1 or 9, characterized in that: Setting the maintenance point includes: determining a number of candidate locations; Based on the fault sequences of the target machine tools and the distances between the corresponding positions of the target machine tools and the candidate positions, an optimal candidate position is determined; and a maintenance point is set at the optimal candidate position.

Images